Multi-Market Systemic Risk Spillover Dataset with Network and Machine Learning Features

This dataset contains 4,316 daily observations from June 5, 2001 to November 14, 2024. It covers twelve major stock market indices across North America, Europe, Asia, and Latin America.

Researchers can use the dataset to examine systemic risk spillovers across global equity markets through network analysis and machine learning. The data combine market-level financial indicators with spillover measures derived from the Diebold and Yilmaz framework, which supports analysis of financial contagion, risk transmission, and predictive modeling.

Each observation includes daily returns, conditional volatility, spillovers transmitted, spillovers received, and net spillover positions. Lagged variables extend these indicators to capture short-term temporal dependence and create a high-dimensional feature set for machine learning models.

The dataset supports research on systemic risk measurement, cross-market contagion, identification of risk transmitters and receivers, and forecasting spillover intensity with supervised learning methods such as Random Forest, XGBoost, and LightGBM. It also allows researchers to compare crisis and non-crisis periods.

By linking multi-market financial indicators with network-based spillover measures, this dataset supports empirical research on financial stability, risk propagation, and early warning systems in financial econometrics, network analysis, and machine learning in finance.